Fast iterative algorithm for the reconstruction of multishot non-cartesian diffusion data.

Area of Science:

• Magnetic Resonance Imaging (MRI)
• Diffusion MRI
• Image Reconstruction

Background:

• Motion artifacts are a significant challenge in multishot non-Cartesian diffusion MRI.
• Current motion-compensated reconstruction methods, like iterative sensitivity-encoded algorithms, suffer from high computational complexity.
• This complexity arises from the need to account for motion-induced phase errors using composite sensitivity functions.

Purpose of the Study:

• To accelerate the motion-compensated iterative reconstruction of multishot non-Cartesian diffusion data.
• To reduce the computational burden associated with existing reconstruction techniques.
• To enable faster and more efficient diffusion MRI acquisition and analysis.

Main Methods:

• A principal component analysis (PCA)-based scheme was employed to reduce the dimensionality of composite sensitivity functions.
• A Toeplitz-based conjugate gradient approach was combined with an augmented Lagrangian optimization scheme.
• This approach facilitates sparse recovery of diffusion data by approximating composite sensitivities with a reduced number of basis functions.

Main Results:

• The proposed simplifications significantly decrease computation time, particularly for under-sampled reconstructions using sparse optimization.
• Achieved reconstruction speeds approximately 9 times faster than conventional methods.
• Effective motion compensation was maintained alongside the accelerated reconstruction.

Conclusions:

• The developed enhancements provide a fast and effective solution for motion-compensated reconstruction of multishot diffusion data.
• The method is applicable to arbitrary k-space trajectories, offering broad utility in diffusion MRI.
• This advancement promises to improve the efficiency and practicality of diffusion MRI studies.